Page 72 - Becoming Metric Wise

62 Becoming Metric-Wise

received several important prizes before it was found out that his publica-
tions were fraudulent. Accusing a colleague of fraudulent behavior, even
if true, is not easy and may harm the career of the whistle blower. Three
case studies illustrating different ways of whistleblowing were shown in
Yong et al. (2013).
An interesting case brought out by Labbe ´ (2010) showed how it is
possible to manipulate automatic rankings and write nonsensical articles,
which however, look quite impressive to an outsider. To do this he used
Scigen (http://pdos.csail.mit.edu/scigen) an automatic generator of arti-
cles, including references, in computer science. These so-called articles
were written by the fictitious author Ike Antkare. By including references
to articles indexed in Google Scholar and by extensive cross-referencing
he was able to give Ike Antkare a Google Scholar h-index (a well-known
bibliometric indicator, see Chapter 6: Journal Citation Analysis and
Chapter 7: Indicators) of 94, making him one of the most visible scientists
in the field.
While doing this he found out that this method had been used by
some persons to get papers accepted at IEEE conferences (Labbe ´ &
Labbe ´, 2013). This finding led Springer and IEEE to remove more than
120 papers from their subscription services, as reported in Van
Noorden (2014).
In view of such cases it would be best if all collaborators on a project
carry out verification tests (in case the collaboration involves lab work)
before submitting the work. Indeed, the fact of being an author carries
responsibilities, concretely with respect of the validity of published work.
Yet, direct verification is not always feasible and there has to be trust in
scientific collaboration, otherwise the science system would break down.
Trust not only applies to collaboration, but also in the publication system
itself. It simply is impossible to redo all experiments that lead to the
results one uses for one’s own work.
Misrepresentation and deliberate over-interpretation of results,
although not data fabrication, are other forms of scientific misconduct.
This also applies to failing to disclose conflicts of interest.

3.3.3 Fake Reviews
A recent form of fraud consists of faking reviews in order to get one’s
papers accepted. This is done as follows: Authors suggest real researchers
as reviewers, but provide fake e-mail addresses. These e-mail addresses